Re: NMUs?

[Samuel Thibault <sthibault@debian.org>]

Svante Signell, le Fri 19 Aug 2011 16:27:45 +0200, a écrit :
> On Fri, 2011-08-19 at 14:56 +0200, Samuel Thibault wrote:
> > Svante Signell, le Fri 19 Aug 2011 14:47:30 +0200, a écrit :
> > > On Fri, 2011-08-19 at 14:36 +0200, Samuel Thibault wrote:
> > > 
> > > > > It could have turned out that this bug hindered a lot of other
> > > > > packages from building.
> > > > 
> > > > Sure, but there's no point in rushing building packages: there's a whole
> > > > year for that. What we need to do is fixing bugs, not hide them.
> > > 
> > > How to address the problem with DMs not willing to package new upstream
> > > versions fixing a lot of bugs, especially including Hurd build problems?
> > > NMU?? This hinders a lot of progress.
> > 
> > Please point a specific case and we'll end up doing an NMU, yes.
> > According to graph-top.txt, there's not so much hindering.
> 
> Two examples:
> 1) file
> Unstable: 5.04-5+b1 (5.04-5: 05 Aug 2010)
> Experimental: 5.04-6: 18 Sep 2010
> Upstream: 5.08: 03 Aug 2011
> Upstream version fixing FTBFS: 5.06: 14 Apr 2011

We already have a file package in the archive, so it's not urging.
NMU-ing a newer upstream version is a tought thing to do. What could be
done, however, is to extract the exact upstream fix and submit to the
bug, and after some time, NMU.

> 2) logrotate
> Unstable: 3.7.8-6: 17 Apr 2010
> Experimental: 3.7.9-1: 22 Jun 2011
> Upstream version fixing FTBFS: 3.8.0: 21 Jun 2011

We also have a logrotate package in the archive, so it's not urging.

> Bug reports:
> #612342: 14 Feb 2011, Partially correct patch, applied to 3.7.9
#613342 actually :)

> #633529: 11 Jul 2011, wishlist about 3.8.0

Again, NMU-ing the newer upstream version is a tought thing. Apparently
Paul Martin is responsive, so he is probably just waiting for the right
time (that is, when he has time to spend on it) to upload the newer
upstream version.

In both case, we already have an installable package in hurd-i386, so
they don't hinder anything and are thus not urging, there is not much
use spending time on them since the fixes are even commited upstream
already. They'll most probably simply arrive with next upstream release
upload, there is no need to spend time to make that happen faster, since
that does not help anything else. If (and only if) no newer upstream
release gets uploaded, we'll have to spend time NMU-ing a patch, but it
takes less time to just bet that the newer upstream will get uploaded
(and it probably will, so it's a safe bet).

Samuel